hanchenye-llvm-project/clang/test/Analysis/NSString.m

407 lines
14 KiB
Mathematica
Raw Normal View History

// RUN: %clang_cc1 -triple i386-apple-darwin10 -analyze -analyzer-experimental-internal-checks -checker-cfref -analyzer-store=region -analyzer-constraints=basic -verify %s
// RUN: %clang_cc1 -triple i386-apple-darwin10 -analyze -analyzer-experimental-internal-checks -checker-cfref -analyzer-store=region -analyzer-constraints=range -verify %s
// RUN: %clang_cc1 -DTEST_64 -triple x86_64-apple-darwin10 -analyze -analyzer-experimental-internal-checks -checker-cfref -analyzer-store=region -analyzer-constraints=basic -verify %s
// RUN: %clang_cc1 -DTEST_64 -triple x86_64-apple-darwin10 -analyze -analyzer-experimental-internal-checks -checker-cfref -analyzer-store=region -analyzer-constraints=range -verify %s
Per an astute observation from Zhongxing Xu, remove a "special case" logic in RegionStoreManager::Retrieve() that was intended to handle conflated uses of pointers as integers. It turns out this isn't needed, and resulted in inconsistent behavior when creating symbolic values on the following test case in 'tests/Analysis/misc-ps.m': typedef struct _BStruct { void *grue; } BStruct; void testB_aux(void *ptr); void testB(BStruct *b) { { int *__gruep__ = ((int *)&((b)->grue)); int __gruev__ = *__gruep__; testB_aux(__gruep__); } { int *__gruep__ = ((int *)&((b)->grue)); int __gruev__ = *__gruep__; if (~0 != __gruev__) {} } } When the code was analyzed with '-arch x86_64', the value assigned to '__gruev__' be would be a symbolic integer, but for '-arch i386' the value assigned to '__gruev__' would be a symbolic region (a blob of memory). With this change the value created is always a symbolic integer. Since the code being removed was added to support analysis of code calling OSAtomicCompareAndSwapXXX(), I also modified 'test/Analysis/NSString.m' to analyze the code in both '-arch i386' and '-arch x86_64', and also added some complementary test cases to test the presence of leaks when using OSAtomicCompareAndSwap32Barrier()/OSAtomicCompareAndSwap64Barrier() instead of just their absence. This code change reveals that previously both RegionStore and BasicStore were handling these cases wrong, and would never cause the analyzer to emit a leak in these cases (false negatives). Now RegionStore gets it right, but BasicStore still gets it wrong (and hence it has been disabled temporarily for this test case). llvm-svn: 84163
2009-10-15 09:40:34 +08:00
// ==-- FIXME: -analyzer-store=basic fails on this file (false negatives). --==
// NOTWORK: %clang_cc1 -triple i386-apple-darwin10 -analyze -analyzer-experimental-internal-checks -checker-cfref -analyzer-store=basic -analyzer-constraints=range -verify %s &&
// NOTWORK: %clang_cc1 -triple i386-apple-darwin10 -analyze -analyzer-experimental-internal-checks -checker-cfref -analyzer-store=basic -analyzer-constraints=basic -verify %s &&
// NOTWORK: %clang_cc1 -DTEST_64 -triple x86_64-apple-darwin10 -analyze -analyzer-experimental-internal-checks -checker-cfref -analyzer-store=basic -analyzer-constraints=basic -verify %s &&
// NOTWORK: %clang_cc1 -DTEST_64 -triple x86_64-apple-darwin10 -analyze -analyzer-experimental-internal-checks -checker-cfref -analyzer-store=basic -analyzer-constraints=range -verify %s
//===----------------------------------------------------------------------===//
// The following code is reduced using delta-debugging from
// Foundation.h (Mac OS X).
//
// It includes the basic definitions for the test cases below.
// Not directly including Foundation.h directly makes this test case
// both svelte and portable to non-Mac platforms.
//===----------------------------------------------------------------------===//
Per an astute observation from Zhongxing Xu, remove a "special case" logic in RegionStoreManager::Retrieve() that was intended to handle conflated uses of pointers as integers. It turns out this isn't needed, and resulted in inconsistent behavior when creating symbolic values on the following test case in 'tests/Analysis/misc-ps.m': typedef struct _BStruct { void *grue; } BStruct; void testB_aux(void *ptr); void testB(BStruct *b) { { int *__gruep__ = ((int *)&((b)->grue)); int __gruev__ = *__gruep__; testB_aux(__gruep__); } { int *__gruep__ = ((int *)&((b)->grue)); int __gruev__ = *__gruep__; if (~0 != __gruev__) {} } } When the code was analyzed with '-arch x86_64', the value assigned to '__gruev__' be would be a symbolic integer, but for '-arch i386' the value assigned to '__gruev__' would be a symbolic region (a blob of memory). With this change the value created is always a symbolic integer. Since the code being removed was added to support analysis of code calling OSAtomicCompareAndSwapXXX(), I also modified 'test/Analysis/NSString.m' to analyze the code in both '-arch i386' and '-arch x86_64', and also added some complementary test cases to test the presence of leaks when using OSAtomicCompareAndSwap32Barrier()/OSAtomicCompareAndSwap64Barrier() instead of just their absence. This code change reveals that previously both RegionStore and BasicStore were handling these cases wrong, and would never cause the analyzer to emit a leak in these cases (false negatives). Now RegionStore gets it right, but BasicStore still gets it wrong (and hence it has been disabled temporarily for this test case). llvm-svn: 84163
2009-10-15 09:40:34 +08:00
#ifdef TEST_64
typedef long long int64_t;
_Bool OSAtomicCompareAndSwap64Barrier( int64_t __oldValue, int64_t __newValue, volatile int64_t *__theValue );
#define COMPARE_SWAP_BARRIER OSAtomicCompareAndSwap64Barrier
typedef int64_t intptr_t;
#else
typedef int int32_t;
Per an astute observation from Zhongxing Xu, remove a "special case" logic in RegionStoreManager::Retrieve() that was intended to handle conflated uses of pointers as integers. It turns out this isn't needed, and resulted in inconsistent behavior when creating symbolic values on the following test case in 'tests/Analysis/misc-ps.m': typedef struct _BStruct { void *grue; } BStruct; void testB_aux(void *ptr); void testB(BStruct *b) { { int *__gruep__ = ((int *)&((b)->grue)); int __gruev__ = *__gruep__; testB_aux(__gruep__); } { int *__gruep__ = ((int *)&((b)->grue)); int __gruev__ = *__gruep__; if (~0 != __gruev__) {} } } When the code was analyzed with '-arch x86_64', the value assigned to '__gruev__' be would be a symbolic integer, but for '-arch i386' the value assigned to '__gruev__' would be a symbolic region (a blob of memory). With this change the value created is always a symbolic integer. Since the code being removed was added to support analysis of code calling OSAtomicCompareAndSwapXXX(), I also modified 'test/Analysis/NSString.m' to analyze the code in both '-arch i386' and '-arch x86_64', and also added some complementary test cases to test the presence of leaks when using OSAtomicCompareAndSwap32Barrier()/OSAtomicCompareAndSwap64Barrier() instead of just their absence. This code change reveals that previously both RegionStore and BasicStore were handling these cases wrong, and would never cause the analyzer to emit a leak in these cases (false negatives). Now RegionStore gets it right, but BasicStore still gets it wrong (and hence it has been disabled temporarily for this test case). llvm-svn: 84163
2009-10-15 09:40:34 +08:00
_Bool OSAtomicCompareAndSwap32Barrier( int32_t __oldValue, int32_t __newValue, volatile int32_t *__theValue );
#define COMPARE_SWAP_BARRIER OSAtomicCompareAndSwap32Barrier
typedef int32_t intptr_t;
#endif
typedef const void * CFTypeRef;
typedef const struct __CFString * CFStringRef;
typedef const struct __CFAllocator * CFAllocatorRef;
extern const CFAllocatorRef kCFAllocatorDefault;
extern CFTypeRef CFRetain(CFTypeRef cf);
void CFRelease(CFTypeRef cf);
typedef const struct __CFDictionary * CFDictionaryRef;
const void *CFDictionaryGetValue(CFDictionaryRef theDict, const void *key);
extern CFStringRef CFStringCreateWithFormat(CFAllocatorRef alloc, CFDictionaryRef formatOptions, CFStringRef format, ...);
typedef signed char BOOL;
typedef int NSInteger;
typedef unsigned int NSUInteger;
@class NSString, Protocol;
extern void NSLog(NSString *format, ...) __attribute__((format(__NSString__, 1, 2)));
typedef NSInteger NSComparisonResult;
typedef struct _NSZone NSZone;
@class NSInvocation, NSMethodSignature, NSCoder, NSString, NSEnumerator;
@protocol NSObject
- (BOOL)isEqual:(id)object;
- (oneway void)release;
- (id)retain;
- (id)autorelease;
@end
@protocol NSCopying
- (id)copyWithZone:(NSZone *)zone;
@end
@protocol NSMutableCopying
- (id)mutableCopyWithZone:(NSZone *)zone;
@end
@protocol NSCoding
- (void)encodeWithCoder:(NSCoder *)aCoder;
@end
@interface NSObject <NSObject> {}
- (id)init;
+ (id)alloc;
@end
extern id NSAllocateObject(Class aClass, NSUInteger extraBytes, NSZone *zone);
typedef struct {} NSFastEnumerationState;
@protocol NSFastEnumeration
- (NSUInteger)countByEnumeratingWithState:(NSFastEnumerationState *)state objects:(id *)stackbuf count:(NSUInteger)len;
@end
@class NSString;
typedef struct _NSRange {} NSRange;
@interface NSArray : NSObject <NSCopying, NSMutableCopying, NSCoding, NSFastEnumeration>
- (NSUInteger)count;
@end
@interface NSMutableArray : NSArray
- (void)addObject:(id)anObject;
- (id)initWithCapacity:(NSUInteger)numItems;
@end
typedef unsigned short unichar;
@class NSData, NSArray, NSDictionary, NSCharacterSet, NSData, NSURL, NSError, NSLocale;
typedef NSUInteger NSStringCompareOptions;
@interface NSString : NSObject <NSCopying, NSMutableCopying, NSCoding> - (NSUInteger)length;
- (NSComparisonResult)compare:(NSString *)string;
- (NSComparisonResult)compare:(NSString *)string options:(NSStringCompareOptions)mask;
- (NSComparisonResult)compare:(NSString *)string options:(NSStringCompareOptions)mask range:(NSRange)compareRange;
- (NSComparisonResult)compare:(NSString *)string options:(NSStringCompareOptions)mask range:(NSRange)compareRange locale:(id)locale;
- (NSComparisonResult)caseInsensitiveCompare:(NSString *)string;
- (NSArray *)componentsSeparatedByCharactersInSet:(NSCharacterSet *)separator;
+ (id)stringWithFormat:(NSString *)format, ... __attribute__((format(__NSString__, 1, 2)));
@end
@interface NSSimpleCString : NSString {} @end
@interface NSConstantString : NSSimpleCString @end
extern void *_NSConstantStringClassReference;
//===----------------------------------------------------------------------===//
// Test cases.
//===----------------------------------------------------------------------===//
NSComparisonResult f1(NSString* s) {
NSString *aString = 0;
return [s compare:aString]; // expected-warning {{Argument to 'NSString' method 'compare:' cannot be nil.}}
}
NSComparisonResult f2(NSString* s) {
NSString *aString = 0;
return [s caseInsensitiveCompare:aString]; // expected-warning {{Argument to 'NSString' method 'caseInsensitiveCompare:' cannot be nil.}}
}
NSComparisonResult f3(NSString* s, NSStringCompareOptions op) {
NSString *aString = 0;
return [s compare:aString options:op]; // expected-warning {{Argument to 'NSString' method 'compare:options:' cannot be nil.}}
}
NSComparisonResult f4(NSString* s, NSStringCompareOptions op, NSRange R) {
NSString *aString = 0;
return [s compare:aString options:op range:R]; // expected-warning {{Argument to 'NSString' method 'compare:options:range:' cannot be nil.}}
}
NSComparisonResult f5(NSString* s, NSStringCompareOptions op, NSRange R) {
NSString *aString = 0;
return [s compare:aString options:op range:R locale:0]; // expected-warning {{Argument to 'NSString' method 'compare:options:range:locale:' cannot be nil.}}
}
NSArray *f6(NSString* s) {
return [s componentsSeparatedByCharactersInSet:0]; // expected-warning {{Argument to 'NSString' method 'componentsSeparatedByCharactersInSet:' cannot be nil.}}
}
NSString* f7(NSString* s1, NSString* s2, NSString* s3) {
NSString* s4 = (NSString*)
CFStringCreateWithFormat(kCFAllocatorDefault, 0, // expected-warning{{leak}}
(CFStringRef) __builtin___CFStringMakeConstantString("%@ %@ (%@)"),
s1, s2, s3);
CFRetain(s4);
return s4;
}
NSMutableArray* f8() {
NSString* s = [[NSString alloc] init];
NSMutableArray* a = [[NSMutableArray alloc] initWithCapacity:2];
[a addObject:s];
[s release]; // no-warning
return a;
}
void f9() {
NSString* s = [[NSString alloc] init];
NSString* q = s;
[s release];
[q release]; // expected-warning {{used after it is released}}
}
NSString* f10() {
static NSString* s = 0;
if (!s) s = [[NSString alloc] init];
return s; // no-warning
}
// Test case for regression reported in <rdar://problem/6452745>.
// Essentially 's' should not be considered allocated on the false branch.
// This exercises the 'EvalAssume' logic in GRTransferFuncs (CFRefCount.cpp).
NSString* f11(CFDictionaryRef dict, const char* key) {
NSString* s = (NSString*) CFDictionaryGetValue(dict, key);
[s retain];
if (s) {
[s release];
}
2009-07-22 02:46:15 +08:00
return 0;
}
// Test case for passing a tracked object by-reference to a function we
// don't understand.
void unknown_function_f12(NSString** s);
void f12() {
NSString *string = [[NSString alloc] init];
unknown_function_f12(&string); // no-warning
}
// Test double release of CFString (PR 4014).
void f13(void) {
CFStringRef ref = CFStringCreateWithFormat(kCFAllocatorDefault, ((void*)0), ((CFStringRef) __builtin___CFStringMakeConstantString ("" "%d" "")), 100);
CFRelease(ref);
CFRelease(ref); // expected-warning{{Reference-counted object is used after it is released}}
}
// Test regular use of -autorelease
@interface TestAutorelease
-(NSString*) getString;
@end
@implementation TestAutorelease
-(NSString*) getString {
NSString *str = [[NSString alloc] init];
return [str autorelease]; // no-warning
}
- (void)m1
{
NSString *s = [[NSString alloc] init]; // expected-warning{{leak}}
[s retain];
[s autorelease];
}
- (void)m2
{
NSString *s = [[[NSString alloc] init] autorelease]; // expected-warning{{leak}}
[s retain];
}
- (void)m3
{
NSString *s = [[[NSString alloc] init] autorelease];
[s retain];
[s autorelease];
}
- (void)m4
{
NSString *s = [[NSString alloc] init]; // expected-warning{{leak}}
[s retain];
}
- (void)m5
{
NSString *s = [[NSString alloc] init];
[s autorelease];
}
@end
@interface C1 : NSObject {}
- (NSString*) getShared;
+ (C1*) sharedInstance;
@end
@implementation C1 : NSObject {}
- (NSString*) getShared {
static NSString* s = 0;
if (!s) s = [[NSString alloc] init];
return s; // no-warning
}
+ (C1 *)sharedInstance {
static C1 *sharedInstance = 0;
if (!sharedInstance) {
sharedInstance = [[C1 alloc] init];
}
return sharedInstance; // no-warning
}
@end
@interface SharedClass : NSObject
+ (id)sharedInstance;
- (id)notShared;
@end
@implementation SharedClass
- (id)_init {
if ((self = [super init])) {
NSLog(@"Bar");
}
return self;
}
- (id)notShared {
2009-02-08 06:55:48 +08:00
return [[SharedClass alloc] _init]; // expected-warning{{leak}}
}
+ (id)sharedInstance {
static SharedClass *_sharedInstance = 0;
if (!_sharedInstance) {
_sharedInstance = [[SharedClass alloc] _init];
}
return _sharedInstance; // no-warning
}
@end
id testSharedClassFromFunction() {
return [[SharedClass alloc] _init]; // no-warning
}
// Test OSCompareAndSwap
_Bool OSAtomicCompareAndSwapPtr( void *__oldValue, void *__newValue, void * volatile *__theValue );
extern BOOL objc_atomicCompareAndSwapPtr(id predicate, id replacement, volatile id *objectLocation);
void testOSCompareAndSwap() {
NSString *old = 0;
NSString *s = [[NSString alloc] init]; // no-warning
if (!OSAtomicCompareAndSwapPtr(0, s, (void**) &old))
[s release];
else
[old release];
}
void testOSCompareAndSwapXXBarrier_local() {
NSString *old = 0;
NSString *s = [[NSString alloc] init]; // no-warning
Per an astute observation from Zhongxing Xu, remove a "special case" logic in RegionStoreManager::Retrieve() that was intended to handle conflated uses of pointers as integers. It turns out this isn't needed, and resulted in inconsistent behavior when creating symbolic values on the following test case in 'tests/Analysis/misc-ps.m': typedef struct _BStruct { void *grue; } BStruct; void testB_aux(void *ptr); void testB(BStruct *b) { { int *__gruep__ = ((int *)&((b)->grue)); int __gruev__ = *__gruep__; testB_aux(__gruep__); } { int *__gruep__ = ((int *)&((b)->grue)); int __gruev__ = *__gruep__; if (~0 != __gruev__) {} } } When the code was analyzed with '-arch x86_64', the value assigned to '__gruev__' be would be a symbolic integer, but for '-arch i386' the value assigned to '__gruev__' would be a symbolic region (a blob of memory). With this change the value created is always a symbolic integer. Since the code being removed was added to support analysis of code calling OSAtomicCompareAndSwapXXX(), I also modified 'test/Analysis/NSString.m' to analyze the code in both '-arch i386' and '-arch x86_64', and also added some complementary test cases to test the presence of leaks when using OSAtomicCompareAndSwap32Barrier()/OSAtomicCompareAndSwap64Barrier() instead of just their absence. This code change reveals that previously both RegionStore and BasicStore were handling these cases wrong, and would never cause the analyzer to emit a leak in these cases (false negatives). Now RegionStore gets it right, but BasicStore still gets it wrong (and hence it has been disabled temporarily for this test case). llvm-svn: 84163
2009-10-15 09:40:34 +08:00
if (!COMPARE_SWAP_BARRIER((intptr_t) 0, (intptr_t) s, (intptr_t*) &old))
[s release];
else
[old release];
}
void testOSCompareAndSwapXXBarrier_local_no_direct_release() {
Per an astute observation from Zhongxing Xu, remove a "special case" logic in RegionStoreManager::Retrieve() that was intended to handle conflated uses of pointers as integers. It turns out this isn't needed, and resulted in inconsistent behavior when creating symbolic values on the following test case in 'tests/Analysis/misc-ps.m': typedef struct _BStruct { void *grue; } BStruct; void testB_aux(void *ptr); void testB(BStruct *b) { { int *__gruep__ = ((int *)&((b)->grue)); int __gruev__ = *__gruep__; testB_aux(__gruep__); } { int *__gruep__ = ((int *)&((b)->grue)); int __gruev__ = *__gruep__; if (~0 != __gruev__) {} } } When the code was analyzed with '-arch x86_64', the value assigned to '__gruev__' be would be a symbolic integer, but for '-arch i386' the value assigned to '__gruev__' would be a symbolic region (a blob of memory). With this change the value created is always a symbolic integer. Since the code being removed was added to support analysis of code calling OSAtomicCompareAndSwapXXX(), I also modified 'test/Analysis/NSString.m' to analyze the code in both '-arch i386' and '-arch x86_64', and also added some complementary test cases to test the presence of leaks when using OSAtomicCompareAndSwap32Barrier()/OSAtomicCompareAndSwap64Barrier() instead of just their absence. This code change reveals that previously both RegionStore and BasicStore were handling these cases wrong, and would never cause the analyzer to emit a leak in these cases (false negatives). Now RegionStore gets it right, but BasicStore still gets it wrong (and hence it has been disabled temporarily for this test case). llvm-svn: 84163
2009-10-15 09:40:34 +08:00
NSString *old = 0;
NSString *s = [[NSString alloc] init]; // no-warning
Per an astute observation from Zhongxing Xu, remove a "special case" logic in RegionStoreManager::Retrieve() that was intended to handle conflated uses of pointers as integers. It turns out this isn't needed, and resulted in inconsistent behavior when creating symbolic values on the following test case in 'tests/Analysis/misc-ps.m': typedef struct _BStruct { void *grue; } BStruct; void testB_aux(void *ptr); void testB(BStruct *b) { { int *__gruep__ = ((int *)&((b)->grue)); int __gruev__ = *__gruep__; testB_aux(__gruep__); } { int *__gruep__ = ((int *)&((b)->grue)); int __gruev__ = *__gruep__; if (~0 != __gruev__) {} } } When the code was analyzed with '-arch x86_64', the value assigned to '__gruev__' be would be a symbolic integer, but for '-arch i386' the value assigned to '__gruev__' would be a symbolic region (a blob of memory). With this change the value created is always a symbolic integer. Since the code being removed was added to support analysis of code calling OSAtomicCompareAndSwapXXX(), I also modified 'test/Analysis/NSString.m' to analyze the code in both '-arch i386' and '-arch x86_64', and also added some complementary test cases to test the presence of leaks when using OSAtomicCompareAndSwap32Barrier()/OSAtomicCompareAndSwap64Barrier() instead of just their absence. This code change reveals that previously both RegionStore and BasicStore were handling these cases wrong, and would never cause the analyzer to emit a leak in these cases (false negatives). Now RegionStore gets it right, but BasicStore still gets it wrong (and hence it has been disabled temporarily for this test case). llvm-svn: 84163
2009-10-15 09:40:34 +08:00
if (!COMPARE_SWAP_BARRIER((intptr_t) 0, (intptr_t) s, (intptr_t*) &old))
return;
else
[old release];
}
int testOSCompareAndSwapXXBarrier_id(Class myclass, id xclass) {
if (COMPARE_SWAP_BARRIER(0, (intptr_t) myclass, (intptr_t*) &xclass))
return 1;
return 0;
Per an astute observation from Zhongxing Xu, remove a "special case" logic in RegionStoreManager::Retrieve() that was intended to handle conflated uses of pointers as integers. It turns out this isn't needed, and resulted in inconsistent behavior when creating symbolic values on the following test case in 'tests/Analysis/misc-ps.m': typedef struct _BStruct { void *grue; } BStruct; void testB_aux(void *ptr); void testB(BStruct *b) { { int *__gruep__ = ((int *)&((b)->grue)); int __gruev__ = *__gruep__; testB_aux(__gruep__); } { int *__gruep__ = ((int *)&((b)->grue)); int __gruev__ = *__gruep__; if (~0 != __gruev__) {} } } When the code was analyzed with '-arch x86_64', the value assigned to '__gruev__' be would be a symbolic integer, but for '-arch i386' the value assigned to '__gruev__' would be a symbolic region (a blob of memory). With this change the value created is always a symbolic integer. Since the code being removed was added to support analysis of code calling OSAtomicCompareAndSwapXXX(), I also modified 'test/Analysis/NSString.m' to analyze the code in both '-arch i386' and '-arch x86_64', and also added some complementary test cases to test the presence of leaks when using OSAtomicCompareAndSwap32Barrier()/OSAtomicCompareAndSwap64Barrier() instead of just their absence. This code change reveals that previously both RegionStore and BasicStore were handling these cases wrong, and would never cause the analyzer to emit a leak in these cases (false negatives). Now RegionStore gets it right, but BasicStore still gets it wrong (and hence it has been disabled temporarily for this test case). llvm-svn: 84163
2009-10-15 09:40:34 +08:00
}
void test_objc_atomicCompareAndSwap_local() {
NSString *old = 0;
NSString *s = [[NSString alloc] init]; // no-warning
if (!objc_atomicCompareAndSwapPtr(0, s, &old))
[s release];
else
[old release];
}
void test_objc_atomicCompareAndSwap_local_no_direct_release() {
Per an astute observation from Zhongxing Xu, remove a "special case" logic in RegionStoreManager::Retrieve() that was intended to handle conflated uses of pointers as integers. It turns out this isn't needed, and resulted in inconsistent behavior when creating symbolic values on the following test case in 'tests/Analysis/misc-ps.m': typedef struct _BStruct { void *grue; } BStruct; void testB_aux(void *ptr); void testB(BStruct *b) { { int *__gruep__ = ((int *)&((b)->grue)); int __gruev__ = *__gruep__; testB_aux(__gruep__); } { int *__gruep__ = ((int *)&((b)->grue)); int __gruev__ = *__gruep__; if (~0 != __gruev__) {} } } When the code was analyzed with '-arch x86_64', the value assigned to '__gruev__' be would be a symbolic integer, but for '-arch i386' the value assigned to '__gruev__' would be a symbolic region (a blob of memory). With this change the value created is always a symbolic integer. Since the code being removed was added to support analysis of code calling OSAtomicCompareAndSwapXXX(), I also modified 'test/Analysis/NSString.m' to analyze the code in both '-arch i386' and '-arch x86_64', and also added some complementary test cases to test the presence of leaks when using OSAtomicCompareAndSwap32Barrier()/OSAtomicCompareAndSwap64Barrier() instead of just their absence. This code change reveals that previously both RegionStore and BasicStore were handling these cases wrong, and would never cause the analyzer to emit a leak in these cases (false negatives). Now RegionStore gets it right, but BasicStore still gets it wrong (and hence it has been disabled temporarily for this test case). llvm-svn: 84163
2009-10-15 09:40:34 +08:00
NSString *old = 0;
NSString *s = [[NSString alloc] init]; // no-warning
Per an astute observation from Zhongxing Xu, remove a "special case" logic in RegionStoreManager::Retrieve() that was intended to handle conflated uses of pointers as integers. It turns out this isn't needed, and resulted in inconsistent behavior when creating symbolic values on the following test case in 'tests/Analysis/misc-ps.m': typedef struct _BStruct { void *grue; } BStruct; void testB_aux(void *ptr); void testB(BStruct *b) { { int *__gruep__ = ((int *)&((b)->grue)); int __gruev__ = *__gruep__; testB_aux(__gruep__); } { int *__gruep__ = ((int *)&((b)->grue)); int __gruev__ = *__gruep__; if (~0 != __gruev__) {} } } When the code was analyzed with '-arch x86_64', the value assigned to '__gruev__' be would be a symbolic integer, but for '-arch i386' the value assigned to '__gruev__' would be a symbolic region (a blob of memory). With this change the value created is always a symbolic integer. Since the code being removed was added to support analysis of code calling OSAtomicCompareAndSwapXXX(), I also modified 'test/Analysis/NSString.m' to analyze the code in both '-arch i386' and '-arch x86_64', and also added some complementary test cases to test the presence of leaks when using OSAtomicCompareAndSwap32Barrier()/OSAtomicCompareAndSwap64Barrier() instead of just their absence. This code change reveals that previously both RegionStore and BasicStore were handling these cases wrong, and would never cause the analyzer to emit a leak in these cases (false negatives). Now RegionStore gets it right, but BasicStore still gets it wrong (and hence it has been disabled temporarily for this test case). llvm-svn: 84163
2009-10-15 09:40:34 +08:00
if (!objc_atomicCompareAndSwapPtr(0, s, &old))
return;
else
[old release];
}
void test_objc_atomicCompareAndSwap_parameter(NSString **old) {
NSString *s = [[NSString alloc] init]; // no-warning
if (!objc_atomicCompareAndSwapPtr(0, s, old))
[s release];
else
[*old release];
}
void test_objc_atomicCompareAndSwap_parameter_no_direct_release(NSString **old) {
NSString *s = [[NSString alloc] init]; // expected-warning{{leak}}
if (!objc_atomicCompareAndSwapPtr(0, s, old))
return;
else
[*old release];
}
Per an astute observation from Zhongxing Xu, remove a "special case" logic in RegionStoreManager::Retrieve() that was intended to handle conflated uses of pointers as integers. It turns out this isn't needed, and resulted in inconsistent behavior when creating symbolic values on the following test case in 'tests/Analysis/misc-ps.m': typedef struct _BStruct { void *grue; } BStruct; void testB_aux(void *ptr); void testB(BStruct *b) { { int *__gruep__ = ((int *)&((b)->grue)); int __gruev__ = *__gruep__; testB_aux(__gruep__); } { int *__gruep__ = ((int *)&((b)->grue)); int __gruev__ = *__gruep__; if (~0 != __gruev__) {} } } When the code was analyzed with '-arch x86_64', the value assigned to '__gruev__' be would be a symbolic integer, but for '-arch i386' the value assigned to '__gruev__' would be a symbolic region (a blob of memory). With this change the value created is always a symbolic integer. Since the code being removed was added to support analysis of code calling OSAtomicCompareAndSwapXXX(), I also modified 'test/Analysis/NSString.m' to analyze the code in both '-arch i386' and '-arch x86_64', and also added some complementary test cases to test the presence of leaks when using OSAtomicCompareAndSwap32Barrier()/OSAtomicCompareAndSwap64Barrier() instead of just their absence. This code change reveals that previously both RegionStore and BasicStore were handling these cases wrong, and would never cause the analyzer to emit a leak in these cases (false negatives). Now RegionStore gets it right, but BasicStore still gets it wrong (and hence it has been disabled temporarily for this test case). llvm-svn: 84163
2009-10-15 09:40:34 +08:00
// Test stringWithFormat (<rdar://problem/6815234>)
void test_stringWithFormat() {
NSString *string = [[NSString stringWithFormat:@"%ld", (long) 100] retain];
[string release];
[string release]; // expected-warning{{Incorrect decrement of the reference count}}
}
// Test isTrackedObjectType().
typedef NSString* WonkyTypedef;
@interface TestIsTracked
+ (WonkyTypedef)newString;
@end
void test_isTrackedObjectType(void) {
NSString *str = [TestIsTracked newString]; // expected-warning{{Potential leak}}
}
// Test isTrackedCFObjectType().
@interface TestIsCFTracked
+ (CFStringRef) badNewCFString;
+ (CFStringRef) newCFString;
@end
@implementation TestIsCFTracked
+ (CFStringRef) newCFString {
return CFStringCreateWithFormat(kCFAllocatorDefault, ((void*)0), ((CFStringRef) __builtin___CFStringMakeConstantString ("" "%d" "")), 100); // no-warning
}
+ (CFStringRef) badNewCFString {
return CFStringCreateWithFormat(kCFAllocatorDefault, ((void*)0), ((CFStringRef) __builtin___CFStringMakeConstantString ("" "%d" "")), 100); // expected-warning{{leak}}
}
// Test @synchronized
void test_synchronized(id x) {
@synchronized(x) {
NSString *string = [[NSString stringWithFormat:@"%ld", (long) 100] retain]; // expected-warning {{leak}}
}
}
@end
void testOSCompareAndSwapXXBarrier_parameter(NSString **old) {
NSString *s = [[NSString alloc] init]; // no-warning
if (!COMPARE_SWAP_BARRIER((intptr_t) 0, (intptr_t) s, (intptr_t*) old))
[s release];
else
[*old release];
}
void testOSCompareAndSwapXXBarrier_parameter_no_direct_release(NSString **old) {
NSString *s = [[NSString alloc] init]; // no-warning
if (!COMPARE_SWAP_BARRIER((intptr_t) 0, (intptr_t) s, (intptr_t*) old))
[s release];
else
return;
}